Would today's Broadband internet speeds have been possible in 2000 if there was a demand?

Factual Questions
1

I first got broadband internet in 2000. It was 1.5 mbps. Pretty slow now, but was a massive improvement over dial up. By 2010, i was around 30 mbps if I remember correctly, and now in 2026 i’m getting close to 500 mbps.

My question is, did they know how to do this in 1999? Like if some benefactor had handed every Telecom a trillion dollars in January 1999 and said “make 500 mbps happen in every household by January 2001”, was the know-how there?

In other words. has the rise in speeds simply been a factor of cost of network upgrades vs. demand? Or has it been more like computer chips, where each generation is necessary to create the following generation like a ladder?

Edit: I’m not saying hand them a trillion and get your best researchers on it. I mean “here’s a trillion to cover your upgrade costs and the army of new technicians you need to hire". Would they have said OK. Or would they have said “we dont even quite know how to do such a thing yet?

2

My understanding is that a lot of what drove what we’d now call high-speed internet was the widespread installation of fiber lines between 2000 and 2010; I believe that DSL lines in the 1990s were often still copper wires.

Fiber optic lines were certainly known technology in 1999, so the know-how might have been there?

3

OC 12 at 601 Mbits/s existed in 1998

I know UW Platteville had that (unknown date) — granted it was for the entire university and not one user…

Brian

4

Yeah, i thought of something similar almost immediately after I submitted this. So yeah it was there. But running that to every house seems ridiculously cumbersome. T1’s at 1.5 mbps have been around since the 60s. I worked at a Telecom (at a low level) in the mid-late 90’s and there were actually a small number of wealthy people who had them for their home.

I was thinking something a little more like what we have now.

5

It’s a little bit of a circular thought exercise. There wasn’t demand because people didn’t think it possible at a reasonable price, so the infrastructure wasn’t built out to support it. Standards have improved so now we have faster tech, but it was never really that we didn’t know how to do it, it was more like the industry didn’t put effort into making it happen faster than demand organically grew. Plus the end equipment (e.g. your PC) was not equipped to handle gigabit+ networking at the time, so even if your ISP developed the plumbing to get it to you (naturally at the higher price for cutting edge performance), nobody would pay it because you couldn’t take advantage of it. So there’s a confluence of reasons that it didn’t exist, but not really hard reasons it couldn’t.

6

From 1999 onwards, gigabit ethernet over CAT5 cable was already possible over twisted pair copper (your standard “internet cable” today that looks like a fat phone line): Gigabit Ethernet - Wikipedia

Apple’s Power Mac G4 and PowerBook G4 already included a 1000 gbps ethernet port a couple years later.

Those speeds were hard to find in a home (or even business) ISP, but you could find them in data centers and maybe between campus networks.

Internet2, for example, was an early network-of-networks between universities, and in 1999, a 2500 Mbit/s connection was already possible (internet2 overview). It got into multiples of gigabits soon after that, and by 2011 reached nearly 9 terabits per second: https://broadbandbreakfast.com/internet2-upgrades-speed-to-8-8-terabits-per-second/

And back in those days, you could also pair multiple ISDN lines together for faster-than-dialup speeds (though not quite T1/T2, etc.). You could also do that with regular dialup modems, combining two or more phone lines and spreading downloads between them (in special software).

So the underlying physical technology was there, but the commercial home availability was not yet. It would take home ISPs much longer to catch up, especially in sparsely-populated countries like the USA (vs much of Asia and parts of Europe). WiFi standards would further lag behind.

7

This site has a simple graphic of average home speeds over time: Average U.S. Internet Speeds Over Time | Ooma

(Too long to paste here, but some extracts…)

  • 14.4 kbps in 1993
  • 56 kbps in 1999
  • 400kbps in 2002
  • 1.1 Mbps in 2005
  • 10 Mbps in 2010
  • 99 Mbps in 2021
  • 214 Mbps in 2025 (though of course you can get faster, and gigabit or 2-gigabit is relatively easy to find today)

And any given part of that timeline, ISPs, data centers, academia, and research labs had faster connections. But home speeds always lagged behind. It’s one thing to deploy a fast link between a few elite institutions; much harder to roll that infrastructure out across millions of households paying $50/mo or less.

8

But what did it cost? Back then there were basically three limiting factors: the availability of a high-speed connection from the nearest “point of presence” (POP) aka “last mile”, the cost of it, and more generally what good it would do you considering the relatively low speeds then of the internet backbone and server connections. It was of course a fine choice for something like an entire university campus.

To put it in perspective, before the broadband explosion, traditional T1 telephone lines offering 1.544 Mbps typically cost between $1000 to $2000 a month and sometimes as much as $3500/mo. I was an early adopter of cable broadband, though I can’t exactly remember the year, but according to Google in this area the service went from pilot to wide deployment around 2000. I do recall that it was theoretically supposed to be 3 Mbps but was much less than that most of the time, and that was when it worked at all. It was notoriously unreliable.

Fortunately, the tech rapidly improved and DOCSIS (Data Over Cable Service Interface Specification) provided a solid reference standard along with reliable modems. DOCSIS 4.0 provides for bidirectional multi-gigabit speeds, and DOCSIS 5.0 is currently under development. I have an ancient obsolete cable modem and a relatively cheap internet service and I reliably get around 700 Mbps, which would have been considered miraculous in 1998!

There’s also DSL, of course, but I’ve never had it and haven’t tracked the technology. I think the general idea is that in most areas the telcos can provide direct fiber to the home and thus compete with and even outperform the cable providers, who themselves may also be able to offer fiber.

9

Google Fiber first offered 1 gigabit speed in 2012. I remember doubting at the time that anyone else would offer those high speeds.

10

DSL by definition is over copper wires, it is still (along with xDSL variants) the only ground-based option in many areas.

I helped implement Siebel CRM for a major pharma c. 1997 and was down at their head office in New Jersey. I remember a network diagram showing the interconnects between their different sites. Most were 56K ISDN connections with one fat T1 line from the Head Office to “Tom’s House”. This was Thomas Siebel, the CEO, who grew the company to 45% of the global CRM market before selling to Oracle.

Needless to say, I was impressed that someone had that kind of service at home while I was still on dialup.

11

Let’s use the DOCSIS standards, as mentioned by @wolfpup as our benchmark of what speeds technology was capable of at different times.

This is just an abridged version of the table from the Wikipedia article.

DOCSIS version Production date Maximum downstream capacity Maximum upstream capacity
1.0 1997 40 Mbit/s 10 Mbit/s
2.0 2002 40 Mbit/s 30 Mbit/s
3.0 2006 1 Gbit/s 200 Mbit/s
3.1 2013 10 Gbit/s 1–2 Gbit/s
4.0 2017 10 Gbit/s 6 Gbit/s

So by this definition, I’d say that the capability to deliver a 40/10 connection over coax to the home existed in 2000. Cable modems and provider equipment capable of this speed should have been available. A modern home computer of the time would have been able to handle those speeds. The just introduced 802.11b WIFI at 11Mbp/s would have been stressed, though.

Also as @wolfpup mentions, getting this speed of connection from all of the homes to the actual Internet would have also required major upgrades at the ISPs. Even if some progressive cable company rolled out full speed DOCSIS 1.0 connections to their customers, they would have needed their own network to be fast enough to handle all of those 40Mbit/s households.

It was only 5-10 years ago in my area when the backhaul speed stopped being a problem. Do you remember back then how your internet connection would slow down at night when everyone got home and went online.

Given a giant pot of money, and a genie with a gun to keep the cable execs from using it for bonuses and stock buybacks, I think such speeds could have been made to be generally available around 2000. The necessary upgrades would have taken a significant amount of time, even with unlimited money.

If things go well, I’ll have a 2Gbp/s synchronous connection by the end of next week, for about $2/day. I’m going to need to invest in some 10G network gear.

12

I’ve got fiber here, 1.5Gb down and 1Gb up. The cable option only gives 50Mb up, in spite of being on DOCSIS 3.1. DOCSIS is still a shared bandwidth system and they would rather give fewer upstream channels in favour of advertising fast downloads. Very disappointing.

13

Of course, even if you had today’s broadband speeds to your house in 2000, there were few, if any, ways to take advantage of it. That was, for example, the reason that Netflix was sending DVDs to people’s houses back then. (Reed Hastings, co-founder of Netflix, claimed once that he always intended Netflix to be a streaming service and the DVD-by-mail service was just a stopgap, although I’m skeptical of that claim.)

14

I have fiber now, but before that I did have DSL – I was an early test customer. I don’t remember the date nor the speed at the time. I know I am < 2 miles from the central office.

ETA: dug up an email from 2003 - I had DSL and I thought it was 756kbps download

Brian

15

I started working for Roku shortly after they spun off from Netflix, and was in some meetings with Hastings. I believe what he says. Besides what I heard in meetings, consider what he named the company. It has nothing to do with discs.

16

ADSL was the a part of the uptick on internet speeds for domestic users. That it could run over exactly the same twisted pair phone lines that were previously only handling 56kb/s of as just plain mind bending.
It was access to cheap enough processing power that made this possible. You need enough signal processing capability to manage the multiple carriers with the crosstalk, reflections and noise in real lines. ADSL modems include what is effectively a time domain reflectometer and adaptive echo cancellation. Similar capabilities are what enables gigabit Ethernet.

Then you need the technology to enable internet providers to build their end. This is paced by access to affordable processing power. Be it switches, routers, or back haul bandwidth.

Optical fibres had long been known as the key enabler for long haul communications. You never needed to update the fibre, each new generation of transceivers brought more bandwidth.

This means rollout was basically waiting for Moore’s Law to catch up. Once you have the ability to fabricate a custom chip solution at scale, everything takes off.

Back in the 90’s high speed internet was a solution in search of a problem. There just weren’t the generally available technologies to allow content generation at scale. Video was a curiosity, with many PCs struggling with even low resolution rendering. An enterprise scale Web server was a Sun E-10000. (Which is why Sun rode Internet boom 1 to mega riches.) Similarly Cisco owned the enterprise network needed. They ran on the edge of what the silicon could provide. And they charged for it. (Which is why they were the other success story of Internet boom 1). Which brings us back to Moore’s Law.

An E-10000 was at the edge of technology. Each ranked in the Top 500 supercomputers. (The story of how Sun came to own the E-10000 is itself an amazing story. They didn’t design it.) Custom silicon enabled a lot of the service side of the Internet.

So, could a trillion dollars have made it happen earlier? Probably, not enough. The research going into enabling technologies was already being funded at high levels. More money would maybe have accelerated things a bit. But we end up with the adage that 9 women can’t produce a baby in one month. Development of silicon fabrication techniques was in its glory years. But the time taken to develop a new node was still years, with lots of technologies all needed to step up. Just getting a trillion dollar project off the ground with the intent of leapfrogging the existing progress would take years. Unlike say Manhattan or Apollo, such a project wouldn’t be starting close to the beginning of technology development, it would have been entering a domain decades old. The easy fast wins were already had.

Alternatively, say you have 100 million households. One trillion dollars is $10,000 dollars each household. Split the money 1:2, $3333to the user end, $6666 to the server end, and subsidise each to cover the much higher cost of technology to achieve the results with a previous generation or two’s capability. Maybe get things at Internet 1 speeds say 5 years earlier. And have a trillion dollars of stranded assets in five years time.

17

This is the demand factor. There was very little in the way of streaming video in 2000 (some, but not much). And operating systems still came on CDs or DVDs. 1.5mbps would have made almost all web page loads instantaneous, and there was that underground mp3 sharing going on, but the demand would take at least a few years to develop.

Perhaps the biggest difference between dial-up and broadband/ISDN/fibre/whatever is the fact that you do not need to log in to the gateway. Having an always-on connection was a huge step forward for a lot of people.

18

I remember going from dial-up to DSL as soon as it was offered (sometime around 2000) and yes, the always-on connection was great. For example, I would be watching a movie and wondering where I’d seen that actor. Easy enough to go to IMDB and find out.

19

For most people, yes, but do not undersell the importance the bump in speed from dialup to DSL. I had always on internet since 1993. Mostly dialup, except the couple of years it was a 19.2 serial connection pushed over two miles of copper wire.

The early DSL I had in 2000 or so was 10-20 times faster than dialup, and that is a huge jump in what can be done. All of the sudden stuff like Net2phone, RealPlayer, and graphic heavy web pages stop being a painful novelty and become usable (with the occasional buffering frustration).

20

I can remember thinking, when my 512Kb ADSL went to 8Mb ADSL2, over the same wire “Why didn’t you just do that in the first place?” Now 70Mb, still the same wire (ish)